As a related-art masking technique which can be used in a monitoring image, an image processing method which recognizes a face or a vehicle number plate for identifying an individual, and hides the recognized area is mainly used. Such a masking technique can hide a detected area, but has a demerit of having to have information on an original image in order to restore the hidden area to the original image.
In Korea, a H. 264 codec-based dynamic privacy masking/unmasking technique has been developed. This technique enables an original image to be restored according to authority, and is a masking technique using a manufactured H. 264 codec. This technique can identify an original image by unmasking a masked image according to whether the corresponding codec is used or not. However, when the codec allowing unmasking is used, the image may be unmasked in a normal video player, and thus there is a problem that personal information may be leaked.
In the related-art image recognition and storage system described above, a front camera installed in a vehicle obtains and records real-time images and voices. In this case, other vehicles or other persons' faces are recorded. When a corresponding video is edited and uploaded on the Internet, there is a problem of infringement of personal information.
Due to the problem of infringement of personal information, in Korea and abroad, there are movements to obligate dashboard cameras currently installed in vehicles to be equipped with a masking function to hide and store vehicle number plates and persons' faces in recording images in the dashboard cameras.
Accordingly, in the future, normal users would not be allowed to identify other persons' faces and others' vehicle number palates even when images are recorded by dashboard cameras installed in their own vehicles. In order to prevent this, vehicle number plates and persons' faces should be automatically masked by using a pattern recognition and image processing technique.
Accordingly, there is a demand for a method for masking a video more efficiently and more securely.